The objective of this application is analysis of data collected from 8/84 through 7/89 in a prospective, longitudinal, randomized trial of secondary prevention of child maltreatment (5ROlMH38373). Because this study is prospective, it was known at the outset that five years would not be sufficient to complete data analysis. Overall aims are: (1) test the predictive validity of the second-edition Maternal History Interview (MHI-II) for prenatal prediction of child maltreatment, (2) test-the effectiveness of an intervention provided from pregnancy through the second postnatal year to prevent maltreatment, and (3) assess an ecological model for causation of child maltreatment. Data will be collected in the current study until funding ends 7/89. This application would allow data analysis using multivariate techniques to complete the above three objectives, to examine prospectively the etiology of child maltreatment, and to conduct studies of medical, social, and developmental outcomes for subject infants. Design of the study included 1,170 prenatal interviews, random assignment of 428 pregnant women designated as "high-risk" by the MHI-II to either intervention or routine medical care at the study hospital, and data collection by persons blind to study hypotheses, group membership, and risk-rating of mothers. The intervention provided comprehensive health care by a multidisciplinary team that emphasized a supportive and nurturing environment through interaction between mothers and team members and through support and education groups conducted by a psychologist. The data set includes the MHI-II, demographic data, maternal and child health information, measures of the mothers' self concepts and well-being; perceptions of their infants; child-rearing styles and attitudes; and observational measures of the home environments. Infant developmental assessments and state child protective service investigations of child maltreatment for study families are also available. Data from a previous prospectively-studied cohort (1974-80) will be used to test the stability of predictive models when compared to the 1984-89 cohort. Structural equation modeling will be used to examine the relationships between the predictive, treatment, and outcome variables and test an ecological model for causation. Event history regression analysis will be used to examine precursors of health and development outcomes of infants and psychological changes in the mothers.